The type and amount of collagen synthesized by tumor cells in vitro appears to be different from that of cultured normal cell counterparts (Stern et al., Cancer Research, in press). We wish to extend these observations to a variety of human sarcomas, and determine whether differentiation in vitro induced by differentiating agents will cause a reversion to more normal patterns of collagen synthesis, and whether these patterns might allow more precise categorization of the tumor's origins. In addition, the analysis will include a search for any new or atypical collagens unique to these tumors. Such collagens, if identified, might be valuable in tumor diagnosis.